Subscribe

Table of Contents

Table of Contents

Clinical

Clinical Profile of Patients Awaiting Keratoplasty in Northern India

Editorial

Letter from the Editor

Featured Article

Corneal Cross-linking for the Treatment of Keratoconus: Laboratory Science, Clinical Effect and the Potential Impact to Eye Banking in the United States

Original Research

Tissue Processing Factors Influencing Descemet Stripping Automated Endothelial Keratoplasty Thickness and Endothelial Outcomes

Challenges in Testing Post-Mortem Blood Samples

Epithelial Desquamation, Regeneration and Renewal in Corneal Graft Prepared for Clinical Use

Corneal Cross-linking for the Treatment of Keratoconus: Laboratory Science, Clinical Effect and the Potential Impact to Eye Banking in the United States

Proceedings

Sinaloa Organ and Tissue Bank: Our Experience in Eye Banking. A 8-Year Review from 2007 through 2015

Corneal Cross-linking for the Treatment of Keratoconus: Laboratory Science, Clinical Effect and the Potential Impact to Eye Banking in the United States

Report

Specular Microscopy Image Quality Scale for Endothelial Cell Density Determination in Donor Cornea Tissue

Specular Microscopy Image Quality Scale for Endothelial Cell Density Determination in Donor Cornea Tissue

Authors

Jianyan Huang, MD, PhD, Tudor C. Tepelus, PhD, Ping Huang, MD, PhD, Adam Fox, CEBT, Andrew J Maxwell, CEBT, Srinivas Sadda, MD, Olivia L. Lee, MD

Abstract

ABSTRACT
Objective: To develop an image quality rating (IQR) scale for
specular microscopy images of donor corneal endothelium and
to evaluate the accuracy and reproducibility of IQR scales based
on inter- and intra-grader variation in endothelial cell density
(ECD) values.

Methods: Using the Konan center method, we performed specular
analysis on 536 specular images of donor corneal endothelium.
All images were assessed independently for image quality
using the IQR grading scale and ECD analysis was performed by
trained specular microscopy graders.

Results: Of the 536 images, 462 (86%) were gradable (excellent,
65 [12%]; good, 230 [43%] fair, 81 [15%]; and poor, 86 [16%].
The remaining 74 (14%) were non-gradable by the image quality
scale. Inter- and intra-grader reproducibility was good (Kappa,
0.77 and 0.88, respectively). When stratified by IQR, the image
quality scale offered comparable high inter- and intra-grader repeatability
(ICC≥0.98) for ECD determination at higher image quality
ratings (excellent/good) and an acceptable confidence level
(ICC>0.8) at lower image quality ratings (fair/poor). The agreement
in ECD measurement (error<5%) is higher in excellent
(96%) and good (>83%) images, but lower in fair (>66%) and
poor (>55%) images, providing further validation of the quality
scale.

Conclusions: For trained graders, image quality is the main
factor affecting ECD reproducibility. This image quality scale
shows good reproducibility for all gradable images. Even when
the IQR score is poor, an acceptable ECD still can be provided.

Key Words: image quality rating, corneal endothelium, donor,
specular microscope

Article PDF

Fullscreen Mode